image_file stringlengths 25 31 | id stringlengths 21 27 | caption stringlengths 9 3.88k | conversations list | first_mention stringlengths 11 5.89k | response stringlengths 24 1.45k | title stringlengths 7 243 | abstract stringlengths 31 3.18k |
|---|---|---|---|---|---|---|---|
1801.01974v2-Figure10-1.png | 1801.01974v2-Figure10-1 | Fig. 10. Average AUC and AUPR versus the number of synthetic ROIs generated with DSFS, 3DMM, and SHBMM according to pose and lighting effects where still-to-video FR system employs either SVM and SRC classifiers on Chokepoint (a, b) and COX-S2V (c, d) databases. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : SDMM+SVMSDMM-CNN+SVM0.750.9SHBMM+SVMDSFS4SVM0.7SDMM+SRCADMM-CNN+SkCSHRMM+NRC0.650.60.50.5040.40.3Baseline0.3519152025Baseline10152025Number of Synthetic Pose and Illumination ROIs per each IndividualNumber of Synthetic Pose and Illumi... | 2) Mixed Pose and Illumination Variations: The performance of still-to-video FR systems is assessed versus the number of synthetic ROIs generated under both pose and lighting effects. Figs. 10a and 10d show average AUC and AUPR obtained by increasing the number of synthetic ROIs used to design SRC and SVM classifiers o... | Question: The graph shows the performance of still-to-video FR systems versus the number of synthetic ROIs generated under both pose and lighting effects. What is the relationship between the number of synthetic ROIs and the performance of the FR system?
Answer: The graph shows that the performance of the FR system im... | Domain-Specific Face Synthesis for Video Face Recognition from a Single
Sample Per Person | The performance of still-to-video FR systems can decline significantly
because faces captured in unconstrained operational domain (OD) over multiple
video cameras have a different underlying data distribution compared to faces
captured under controlled conditions in the enrollment domain (ED) with a still
camera. This ... |
1801.05671v1-Figure9-1.png | 1801.05671v1-Figure9-1 | Figure 9: Static target and PPS modulation: human hand vs. head. See Fig. 8 for explanation of individual panels and text for details. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : distance to end-effector (m)10.8-0.60.40.2distance to elbow (m)1.0.80.60.40.2end-effector error (m)0.080.060.04 -0.02101520253035time (s)5., distance, to, end-effector, (m), 1, 0.8-, 0.6, 0.4, 0.2, distance, to, elbow, (m), 1., 0.8, 0... | The setup for this experiment is similar to Section 3.1. The main difference is that we assigned different relevance for the different human body parts, as specified in Section 2.4. This directly relates to a realistic human-robot interaction in which the safety of some body parts (e.g., head) should be guaranteed with... | Question: How does the robot's behavior change when the PPS is modulated for different body parts?
Answer: The robot's behavior changes when the PPS is modulated for different body parts. When the PPS is modulated for the human hand, the robot can get closer to the hand without triggering a strong response of the PPS ... | Compact Real-time avoidance on a Humanoid Robot for Human-robot
Interaction | With robots leaving factories and entering less controlled domains, possibly
sharing the space with humans, safety is paramount and multimodal awareness of
the body surface and the surrounding environment is fundamental. Taking
inspiration from peripersonal space representations in humans, we present a
framework on a h... |
2004.06653v1-Figure9-1.png | 2004.06653v1-Figure9-1 | Figure 9: The effect of θ. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 3500IRJQ UP +IRQ_UPIRJQIRQ Eаccuracyrecall Xасcuracy3000rеcall1.01.025001000IRJO UPIRQ_UPIRJQIRQ E20000.80.8800中日日-日0.60.6150060010000.40.4400200 **5000.20.223米0.00.00.30.5theta0.10.70.10.30.5theta0.70.30.5theta0.70.30.5theta0.10.90.9... | Figure 9 shows the effect of precision threshold θ. On the datasets of MPG and MPG5, the runtime of pruned algorithms (IRQ and IRJQ) decreases with an increasing threshold θ and unpruned algorithms (IRQ UP and IRJQ UP) keep the approximately same value, which verify the efficiency of our pruning strategies. The recall ... | Question: What is the effect of the precision threshold θ on the runtime and accuracy of the pruned and unpruned algorithms?
Answer: The runtime of pruned algorithms (IRQ and IRJQ) decreases with an increasing threshold θ and unpruned algorithms (IRQ UP and IRJQ UP) keep the approximately same value, which verify the ... | Efficient Suspected Infected Crowds Detection Based on Spatio-Temporal
Trajectories | Virus transmission from person to person is an emergency event facing the
global public. Early detection and isolation of potentially susceptible crowds
can effectively control the epidemic of its disease. Existing metrics can not
correctly address the infected rate on trajectories. To solve this problem, we
propose a ... |
1803.06598v1-Figure8-1.png | 1803.06598v1-Figure8-1 | Figure 8: Performances of different Gaussian sampling in the 300-W public testing set. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : Performance of Different Gaussian Distribution0.0640.0620.0600.0580.0560.0540.0520.0500.048.00.10.20.30.40.50.6 0.70.80.91.0Gaussian Distribution Parametar: oNormalized Mean Eror, Performance, of, Different, Gaussian, Distribution, 0.... | Effect of Gaussian sampling space parameters. As one of the most important processes, random sampling space significantly affects the final robustness and accuracy. As shown in Figure 8, the NME results are presented by varying the standard deviation σ of Gaussian sampling. Appropriate values lead to promising performa... | Question: What is the effect of the Gaussian sampling space parameters on the performance of the proposed method?
Answer: The Gaussian sampling space is one of the most important processes in the proposed method. The standard deviation σ of Gaussian sampling significantly affects the final robustness and accuracy. As ... | Facial Landmarks Detection by Self-Iterative Regression based
Landmarks-Attention Network | Cascaded Regression (CR) based methods have been proposed to solve facial
landmarks detection problem, which learn a series of descent directions by
multiple cascaded regressors separately trained in coarse and fine stages. They
outperform the traditional gradient descent based methods in both accuracy and
running spee... |
2004.04071v2-Figure7-1.png | 2004.04071v2-Figure7-1 | Figure 7: behanviour of variance (left) and cost (right) of the multilevel difference ∆Q`,`−1 with level `. The homogeneous case (a = 0) is indicated by the dashed line ( ), and the heterogeneous case (a 0) is indicated by the full line ( ). | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 1/R(x)1/R(x), 1/R(x), 1/R(x) Figure 7: behanviour of variance (left) and cost (right) of the multilevel difference ∆Q`,`−1 with level `. The homogeneous case (a = 0) is indicated by the dashed line ( ), and the heterogeneous case (a 0... | The effect of the level selection strategy can be visualized by inspecting the variances V` and V`,`−1, ` = 0, 1, . . . , L and costs C` and C`,`−1, ` = 0, 1, . . . , L, see Figures 5–6. These figures indicate a similar behanviour of the variance and cost of the multilevel difference for all parameter combinations, see... | Question: What is the effect of the level selection strategy on the variance and cost of the multilevel difference?
Answer: The effect of the level selection strategy can be visualized by inspecting the variances V` and V`,`−1, ` = 0, 1, . . . , L and costs C` and C`,`−1, ` = 0, 1, . . . , L, see Figures 5–6. These fi... | Multilevel Asymptotic-Preserving Monte Carlo for Particle Simulations | We develop a novel multilevel asymptotic-preserving Monte Carlo method,
called Multilevel Kinetic-Diffusion Monte Carlo (ML-KDMC), for simulating the
kinetic Boltzmann transport equation with a Bhatnagar-Gross-Krook (BGK)
collision operator. This equation occurs, for instance, in mathematical models
of the neutral part... |
1603.08982v1-Figure2-1.png | 1603.08982v1-Figure2-1 | Fig. 2: MSE vs. SNR under K-distributed noise. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 10°CMLE* IMAPE, one iterationO IMAPE, two iterations- CRB101010101015SNR (dB)2025MSE, 10°, CMLE, *, IMAPE,, one, iteration, O, IMAPE,, two, iterations, -, CRB, 10, 10, 10, 10, 10, 15, SNR, (dB), 20, 25, MSE Fig. 2: MSE vs. SNR under K... | In Figs. 1 and 2, we plot the mean square errors (MSEs) of the estimation of θ under the SIRP noise versus the SNR by implementing our proposed IMLE and IMAPE algorithms, respectively. In Fig. 1 the noise is t-distributed, and in Fig. 2, K-distributed. For | Question: How does the MSE of the estimation of θ change with the SNR under K-distributed noise?
Answer: The MSE of the estimation of θ decreases with the SNR under K-distributed noise. This is because the SNR is a measure of the signal power relative to the noise power, and as the SNR increases, the signal power beco... | Maximum Likelihood and Maximum A Posteriori Direction-of-Arrival
Estimation in the Presence of SIRP Noise | The maximum likelihood (ML) and maximum a posteriori (MAP) estimation
techniques are widely used to address the direction-of-arrival (DOA) estimation
problems, an important topic in sensor array processing. Conventionally the ML
estimators in the DOA estimation context assume the sensor noise to follow a
Gaussian distr... |
1411.4940v2-Figure9-1.png | 1411.4940v2-Figure9-1 | Fig. 9: Vary number of objects | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : dVP-B+ mVP-TPR* -O dVP-TPR* -O SP-TPR*+mVP-B+ SP-B**B-* TPR60404020100K200K300K400KSOOK100K200K300K400K5OOK100K200K300K400K5OOKNumber of objectsNumber of objectsNumber of objects(a) Throughput(b) Range query(c) kNN queryResponse time ... | In the next experiment, we vary the number of moving objects from 100K to 500K. Figure 9 shows the results about throughput, range query and kNN query. We can see that when the number of objects increases, throughput decreases, query response time increases for both range queries and kNN queries. Moreover, Bx-trees enj... | Question: How does the throughput and query response time change with the number of objects?
Answer: The throughput decreases and the query response time increases as the number of objects increases. This is because the number of objects increases, the number of nodes in the index increases, and the index becomes more... | Speed Partitioning for Indexing Moving Objects | Indexing moving objects has been extensively studied in the past decades.
Moving objects, such as vehicles and mobile device users, usually exhibit some
patterns on their velocities, which can be utilized for velocity-based
partitioning to improve performance of the indexes. Existing velocity-based
partitioning techniq... |
2005.11194v2-Figure4-1.png | 2005.11194v2-Figure4-1 | Figure 4: The training history of our neural network trained to predict log(arsenic) in stream sediments from terrain texture. The vertical dashed line marks the best epoch, for which the mean-square-error (MSE) on held out test data is lowest. The weights at this best epoch are the ones that are kept for subsequent us... | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 2.0-1.5 -dataset1.0-trainingtesting0.5-0.0-50100150200250ерochMSE, 2.0-, 1.5, -, dataset, 1.0-, training, testing, 0.5-, 0.0-, 50, 100, 150, 200, 250, ерoch, MSE Figure 4: The training history of our neural network trained to predict ... | To train the neural network, for each element we split our dataset into 10 folds at random, and trained using 9 of them, while monitoring the mean-squared-error (MSE) of the neural network’s predictions on the 10th fold to ensure that we did not overfit. We trained the neural network using the ADAM optimiser, and a bat... | Question: What is the best epoch for the neural network to predict log(arsenic) in stream sediments from terrain texture?
Answer: The best epoch for the neural network to predict log(arsenic) in stream sediments from terrain texture is 200. This is because the mean-square-error (MSE) on held out test data is lowest at... | Deep covariate-learning: optimising information extraction from terrain
texture for geostatistical modelling applications | Where data is available, it is desirable in geostatistical modelling to make
use of additional covariates, for example terrain data, in order to improve
prediction accuracy in the modelling task. While elevation itself may be
important, additional explanatory power for any given problem can be sought
(but not necessari... |
1103.5426v3-Figure7-1.png | 1103.5426v3-Figure7-1 | Fig. 7. Numerical analysis: gap between achievable scheme and the outerbounds in Corollary 1 at (a) SNR = 20dB, (b) SNR = 40dB, and (c) SNR = 60dB for CFB = 10 bits. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : SNR = 20DB0.5INR (dB)(a)SNR = 40dB1050INŘ (dB)(b)SNR = 40dB1.50.5INR (dB)(c)gap (bits/sec/Hz)gap (bits/sec/Hz)gap (bits/sec/Hz)%24, SNR, =, 20DB, 0.5, INR, (dB), (a), SNR, =, 40dB, 10, 50, INŘ, (dB), (b), SNR, =, 40dB, 1.5, 0.5, INR, ... | (i) k : k ∈ {1, 2}, i ∈ {1, 2, 3, 4}). In general, one can solve an optimization problem to find the optimal choice of power level assignments that maximizes the achievable sum-rate. We have performed numerical analysis for this optimization problem. Figure 7 shows the gap between our proposed achievable scheme and the... | Question: What is the relationship between the gap between the achievable scheme and the outer-bounds in Corollary 1, and the SNR and INR values?
Answer: The graph shows the gap between the achievable scheme and the outer-bounds in Corollary 1 at (a) SNR = 20dB, (b) SNR = 40dB, and (c) SNR = 60dB, for CFB = 10 bits. T... | Interference Channels with Rate-Limited Feedback | We consider the two-user interference channel with rate-limited feedback.
Related prior works focus on the case where feedback links have infinite
capacity, while no research has been done for the rate-limited feedback
problem. Several new challenges arise due to the capacity limitations of the
feedback links, both in ... |
1301.0216v1-Figure9-1.png | 1301.0216v1-Figure9-1 | Figure 9: Percentage of groups for which a timetable was found as a function of group size. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 100S1: Scotland (trains)S2: Scotland (trains, coaches)S3: Central UK (trains)90S4: Central UK (trains, coaches)80S5: South and central UK (trains)7060504030101.356 78group size number of agents2.20groups with timetable [%), 100, S1:, ... | The timetabling phase is of course much more problematic. Figure 9 shows the percentage of groups for which a timetable was found, as a function of group size. In order to create this graph, number of groups with assigned timetable and total number of groups identified was counted for every size of the group. There are... | Question: What is the relationship between the percentage of groups for which a timetable was found and the group size?
Answer: The graph shows that as the group size increases, the percentage of groups for which a timetable was found decreases. This is likely because as the group size increases, it becomes more diffi... | Applying Strategic Multiagent Planning to Real-World Travel Sharing
Problems | Travel sharing, i.e., the problem of finding parts of routes which can be
shared by several travellers with different points of departure and
destinations, is a complex multiagent problem that requires taking into account
individual agents' preferences to come up with mutually acceptable joint plans.
In this paper, we ... |
1610.06815v1-Figure4-1.png | 1610.06815v1-Figure4-1 | Fig. 4. Learning error for different learning rates. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 350Learning Rate300· 0.0050.01250- 0.01820015010050100200300400500Number of IterationError, 350, Learning, Rate, 300, ·, 0.005, 0.01, 250, -, 0.018, 200, 150, 100, 50, 100, 200, 300, 400, 500, Number, of, Iteration, Error Fig. 4. Lear... | 1) Effects of Momentum and Learning Rate in Pretraining: A good combination of learning rate and momentum is crucial for the convergence of RBM learning. Usually, a small value for the learning rate and a large value for the momentum are helpful for the convergence. Fig. 4 shows reconstruction errors for a Gaussian–Ber... | Question: What is the relationship between the learning rate and the reconstruction error?
Answer: The graph shows the reconstruction error for a Gaussian–Bernoulli RBM with a structure of 312-600. We set the momentum value as 0.9 and the learning rate values as 0.005, 0.01, and 0.018, respectively. It can be observed... | Deep Models for Engagement Assessment With Scarce Label Information | Task engagement is defined as loadings on energetic arousal (affect), task
motivation, and concentration (cognition). It is usually challenging and
expensive to label cognitive state data, and traditional computational models
trained with limited label information for engagement assessment do not perform
well because o... |
1904.07061v1-Figure13-1.png | 1904.07061v1-Figure13-1 | Figure 13. Performance measurements for fused FMM traversals normalized to the unfused ones for different number of points. Number of points on the x axis and normalized measurement on the y axis. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 1.20.920.910.80.780.60.40.21.0E+051.0E+061.0E+071.0E+08+runtime +Ļ2 misses -=[3 misses *instructions, 1.2, 0.92, 0.91, 0.8, 0.78, 0.6, 0.4, 0.2, 1.0E+05, 1.0E+06, 1.0E+07, 1.0E+08, +runtime, +Ļ2, misses, -=[3, misses, *instructions Fi... | The fast multipole method (FMM) is a numerical technique used in evaluating pairwise interactions between large number of points distributed in a space (e.g., Long-ranged forces in the n-body problem, computation of gravitational potential, and computational electromagnetic problems) [12]. In this case study, we reimpl... | Question: What is the relationship between the number of points and the performance of the fused traversals?
Answer: The graph shows that the performance of the fused traversals increases with the number of points. This is because the fused traversals are able to fully fuse the two passes and yield a performance impro... | Sound, Fine-Grained Traversal Fusion for Heterogeneous Trees - Extended
Version | Applications in many domains are based on a series of traversals of tree
structures, and fusing these traversals together to reduce the total number of
passes over the tree is a common, important optimization technique. In
applications such as compilers and render trees, these trees are heterogeneous:
different nodes o... |
1904.05777v2-Figure6-1.png | 1904.05777v2-Figure6-1 | Figure 6. (a) Comparison of the MSE of various CS reconstruction algorithms in the presence of correlated measurement matrices with k = 5. The signals to be retrieved have density ρ = 0.5. (b) Semi-logarithmic plot of the reconstruction error of Basis Pursuit, SL0 and EP. (c) Comparison of the elapsed running time of t... | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 0.6+ AMP-- Basis Pursuit0.5* COSAMP0.4솔 EMBP* OMP• ROMPSubspacePursuit* SLO0.30.2+ EP0.10.00.70.80.9(a)10°.10'+ AMP|-X Basis Pursuit* COSAMP10-510010-10A EMBP101.* OMP10-15ROMP+ SubspacePursuit* SLOBasis Pursuit10 2,10-20SLOЕРЕР-25100... | random gaussian i.i.d. signals of length N = 100 and as many random correlated sensing matrices, with k = 5. For any given pair of signal w and measurement matrix F, we attempted to recover the original signal by means of EP and of the algorithms included in KL1p. The results are presented in figure 6. As we can see in... | Question: What is the relationship between the MSE of various CS reconstruction algorithms in the presence of correlated measurement matrices with k = 5?
Answer: The graph shows that the MSE of EP is the lowest among all the algorithms, and it is the only algorithm that exhibits an incorrect-correct reconstruction pha... | Compressed sensing reconstruction using Expectation Propagation | Many interesting problems in fields ranging from telecommunications to
computational biology can be formalized in terms of large underdetermined
systems of linear equations with additional constraints or regularizers. One of
the most studied ones, the Compressed Sensing problem (CS), consists in finding
the solution wi... |
1804.02167v1-Figure3-1.png | 1804.02167v1-Figure3-1 | Fig. 3. RMSE in concentration of the MAP-MHE state estimator as a function of time, for a random network of 5 threshold sensors. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 200400600Time (s)80010001200RMSE in concentration (g/m)4., 200, 400, 600, Time, (s), 800, 1000, 1200, RMSE, in, concentration, (g/m), 4. Fig. 3. RMSE in concentration of the MAP-MHE state estimator as a function of time, for a random ... | Fig. 3 shows the performance of the novel MH-MAP state estimator implemented in MATLAB, in terms of Root Mean Square Error (RMSE) of the estimated concentration field, i.e.: | Question: What is the relationship between the RMSE in concentration of the MAP-MHE state estimator and the time, for a random network of 5 threshold sensors?
Answer: The graph shows the RMSE in concentration of the MAP-MHE state estimator as a function of time, for a random network of 5 threshold sensors. The RMSE de... | MAP moving horizon state estimation with binary measurements | The paper addresses state estimation for discrete-time systems with binary
(threshold) measurements by following a Maximum A posteriori Probability (MAP)
approach and exploiting a Moving Horizon (MH) approximation of the MAP
cost-function. It is shown that, for a linear system and noise distributions
with log-concave p... |
1910.11851v1-Figure6-1.png | 1910.11851v1-Figure6-1 | Figure 6: Probability density functions of node popularities pi from empirical data, the greedy model, and probabilistic model. Both the greedy pi and probabilistic pi match the data well. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : datamodel greedymodel probyangpu10-s10-410-3PDF, data, model, greedy, model, prob, yangpu, 10-s, 10-4, 10-3, PDF Figure 6: Probability density functions of node popularities pi from empirical data, the greedy model, and probabilistic ... | each dataset we computed both the empirical pi along with the ‘model’ pi, those produced by a fleet of RNN-taxis. The latter were found by generating NT trajectories of length 〈L〉 where the empirical NT and 〈L〉 were used (i.e. we calculated NT and 〈L〉, the mean trip length, from the datasets). Trajectories were generat... | Question: What is the relationship between the probability density functions of node popularities pi from empirical data, the greedy model, and probabilistic model?
Answer: The graph shows the probability density functions of node popularities pi from empirical data, the greedy model, and probabilistic model. Both the... | Modeling vehicular mobility patterns using recurrent neural networks | Data on vehicular mobility patterns have proved useful in many contexts. Yet
generative models which accurately reproduce these mobility patterns are
scarce. Here, we explore if recurrent neural networks can cure this scarcity.
By training networks on taxi from NYC and Shanghai, and personal cars from
Michigan, we show... |
1803.09588v1-Figure7-1.png | 1803.09588v1-Figure7-1 | Fig. 7. Regression snap shots obtained at epoch 1, 5, 10 and at convergence of the regular probe nets. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 10080604020+ Epoch 1, R2 = 0.896* Epoch 5, R? = 0.962I Epoch 10, R² = 0.960Converged, R2 = 0.95220406080100probe net accuracy++ResNet-20 accuracy, 100, 80, 60, 40, 20, +, Epoch, 1,, R2, =, 0.896, *, Epoch, 5,, R?, =, 0.962, I, Epoch, ... | it is trained on all datasets. Within a few epochs the regression quality reaches a saturation at about R2 = 0.95. The mean accuracy difference between the probe nets and the reference ResNets (trained till convergence) is further decreased, meaning that the probe nets are not yet converged and are still increasing the... | Question: What is the relationship between the accuracy of the probe nets and the ResNet-20?
Answer: The probe nets are not yet converged and are still increasing their own classification performance. | Efficient Image Dataset Classification Difficulty Estimation for
Predicting Deep-Learning Accuracy | In the deep-learning community new algorithms are published at an incredible
pace. Therefore, solving an image classification problem for new datasets
becomes a challenging task, as it requires to re-evaluate published algorithms
and their different configurations in order to find a close to optimal
classifier. To faci... |
1304.4928v1-Figure1-1.png | 1304.4928v1-Figure1-1 | Figure 1. Higher cost-to-benefit ratios in the snowdrift game lead to lower levels of cooperation. Depicted are results of an economic experiment, as obtained in the absence of punishment. Panels (a) and (b) show the frequency of cooperation fc and the average payoff per period in dependence on the cost-to-benefit rati... | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 1.0(a)° 0.50.0(b)3.01.50.00.00.20.40.60.81.0cost to benefit ratio raverage payoff, 1.0, (a), °, 0.5, 0.0, (b), 3.0, 1.5, 0.0, 0.0, 0.2, 0.4, 0.6, 0.8, 1.0, cost, to, benefit, ratio, r, average, payoff Figure 1. Higher cost-to-benefit ... | for r = 0.2 and r = 0.8, other four sessions for other values of r). In each session 20 undergraduate students were randomly allocated to four groups of five subjects playing snowdrift games in the Computer Lab for Behavior Games [for further information see the Methods and Fig. S1 (Supplementary Information)]. As it i... | Question: What is the relationship between the cost-to-benefit ratio and the level of cooperation?
Answer: The graph shows that the level of cooperation decreases with increasing of the cost-to-benefit ratio. This is because the best strategy for individuals is to defect if the opponent adopts cooperation. Consequentl... | If cooperation is likely punish mildly: Insights from economic
experiments based on the snowdrift game | Punishment may deter antisocial behavior. Yet to punish is costly, and the
costs often do not offset the gains that are due to elevated levels of
cooperation. However, the effectiveness of punishment depends not only on how
costly it is, but also on the circumstances defining the social dilemma. Using
the snowdrift gam... |
2007.11503v2-Figure22-1.png | 2007.11503v2-Figure22-1 | Figure 22: LLAMA provides improvements to channel capacity and power in a reflective configuration. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : -200.6-300.5-400.4-50H0.3-600.2With metasurfaceWithout metasurface-With metasurface-70Without metasurface-0.12430364248546066Tx-Metasurface distance (cm)Received signalpower (dBm)Capacity (Mbps/Hz), -20, 0.6, -30, 0.5, -40, 0.4, -50, ... | In addition to evaluating transmissive configurations, we also look at LLAMA’s effect on reflected signals. We place the transmitter and receiver on the same side of the metasurface, and separate the transmitter and receiver by 70 cm. We perform experiments at different Tx-metasurface distances by moving the metasurfac... | Question: How does the metasurface affect the received signal power and channel capacity in a reflective configuration?
Answer: The metasurface improves the received signal power and channel capacity in a reflective configuration. The improvement is more significant at shorter Tx-metasurface distances. | Pushing the Physical Limits of IoT Devices with Programmable
Metasurfaces | Small, low-cost IoT devices are typically equipped with only a single,
low-quality antenna, significantly limiting communication range and link
quality. In particular, these antennas are typically linearly polarized and
therefore susceptible to polarization mismatch, which can easily cause 10-15
dBm of link loss on com... |
1904.01697v4-Figure10-1.png | 1904.01697v4-Figure10-1 | Fig. 10: Comparing BER for different values of diffusion coefficient dr | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : Symbol 0 is sentSymbol 1 is sent0.140.40.350.130.320.120.2550.110.25 0.10.150.090.1-0.080.050.10.20.30.40.50.60.70.80.91.0.10.20.30.40.50.60.70.80.9dr (Diffusion rate)dr (Diffusion rate)(b)Bit Error Rate, Symbol, 0, is, sent, Symbol, ... | The aim of this experiment is to further study the impact of diffusion of the receiver species on BER. We choose M “ 10 and P “ 2. We vary dr µm2s´1 from 0 to 1 with an increment of 0.1. We used the BER at time 2.5 sec for comparison. Figs. 10a and 10b show how BER varies with dr for, respectively, Symbols 0 and 1. It ... | Question: What is the relationship between the diffusion rate and the bit error rate for symbols 0 and 1?
Answer: The graph shows that the bit error rate increases monotonically with the diffusion rate for both symbols 0 and 1. This is because as the diffusion rate increases, the receiver species diffuses more quickly... | Using spatial partitioning to reduce the bit error rate of
diffusion-based molecular communications | This work builds on our earlier work on designing demodulators for
diffusion-based molecular communications using a Markovian approach. The
demodulation filters take the form of an ordinary differential equation (ODE)
which computes the log-posteriori probability of observing a transmission
symbol given the continuous ... |
1801.05500v1-Figure8-1.png | 1801.05500v1-Figure8-1 | Fig. 8: Effect of the ground network densification on the average (a) wireless latency per UAV and (b) rate per ground UE for different utility functions and for various altitudes of the UAVs. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : (a)(b)30Altitude=D120mAltitude=D180mAltitude=240m0.925200.70.615Altitude=120mAltitude=D180mAltitude=240m'0.510100.4101520253015202530Number of ground BSsNumber of ground BSsAverage wireless latency per UAV (ms)Average rate per ground ... | Fig. 8 investigates the (a) wireless latency per UAV and (b) rate per ground UE for different values of the UAVs altitude and as a function of the number of BSs. From this figure, we can see that as the UAV altitude increases and/or the ground network becomes denser, the average wireless latency per UAV increases. For ... | Question: How does the average wireless latency per UAV and the average rate per ground UE change with the number of ground BSs and the altitude of the UAVs?
Answer: The average wireless latency per UAV increases as the UAV altitude increases and/or the ground network becomes denser. The average rate per ground UE dec... | Cellular-Connected UAVs over 5G: Deep Reinforcement Learning for
Interference Management | In this paper, an interference-aware path planning scheme for a network of
cellular-connected unmanned aerial vehicles (UAVs) is proposed. In particular,
each UAV aims at achieving a tradeoff between maximizing energy efficiency and
minimizing both wireless latency and the interference level caused on the
ground networ... |
1504.07880v2-Figure5-1.png | 1504.07880v2-Figure5-1 | Figure 5. Influence of candidate paths on usual RCSP with k = 1 and k = 10 constraint | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : k = 1102k =10A*102101+ cor.101e dom.10010010-110-110-210-2A*10-3+ Cor.10-3e dom.102103104105Vertices102103104105VerticesCPU time (s)CPU time (s)||, k, =, 1, 102, k, =, 10, A*, 102, 101, +, cor., 101, e, dom., 100, 100, 10-1, 10-1, 10-... | bound test, is is always more interesting to use the key c(xP ⊕ bv) than the key c(xP ). In practice, the key has an influence only on weakly constrained problems with k = 1, where the keys identify which partial solutions are the most promising. On more constrained problems with k = 10, we are not able to guess which ... | Question: How does the use of candidate paths affect the performance of the algorithms on the wide grid instances with k = 1 and k = 10 constraints?
Answer: The graph shows that the use of candidate paths is not interesting on instances that can be solved to optimality, as the decrease in cUBod is not large enough to ... | Algorithms for Non-Linear and Stochastic Resource Constrained Shortest
Paths | Resource constrained shortest path problems are usually solved thanks to a
smart enumeration of all the non-dominated paths. Recent improvements of these
enumeration algorithms rely on the use of bounds on path resources to discard
partial solutions. The quality of the bounds determines the performance of the
algorithm... |
1110.2739v1-Figure8-1.png | 1110.2739v1-Figure8-1 | Figure 8: The curves for 3-XOR-SAT, 3-Max-rank and (a,3)-QXOR-SAT for n=1000. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 1.0.90.80.70.60.50.40.3#clauses/#exvars = 0.9183-XOR-SAT for n=1k--e--0.23-Max-rank for n=1k----e----(1,3)-QXOR-SAT for m=n-- +0.10.80.850.90.951#clauses/#exvars, 1., 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, 0.3, #clauses/#exvars, =, 0.918, 3-XO... | It can be observed from the experimental results shown in Figure 8 that, contrary to what has been observed in the previous section, the three smooth lines connecting the consecutive points and corresponding to the transition of the three properties 3-Max-rank, (a,3)-QXOR-SAT and 3-XOR-SAT are difficult to distinguish.... | Question: What is the relationship between the three properties 3-Max-rank, (a,3)-QXOR-SAT and 3-XOR-SAT?
Answer: The three properties 3-Max-rank, (a,3)-QXOR-SAT and 3-XOR-SAT have a sharp phase transition at the critical value c3 ≈ 0.918. | Phase Transition for Random Quantified XOR-Formulas | The QXORSAT problem is the quantified version of the satisfiability problem
XORSAT in which the connective exclusive-or is used instead of the usual or. We
study the phase transition associated with random QXORSAT instances. We give a
description of this phase transition in the case of one alternation of
quantifiers, t... |
1006.4386v1-Figure3-1.png | 1006.4386v1-Figure3-1 | Fig. 3. DF Second-hop secrecy rate vs. the total relay transmit powerPT for different cases. Eavesdropper has a stronger channel. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 109令GG487中一ロー G日o00d4Total Power constrainte - individual power constraint, SDR approachA individual power constraint, SOCP approch3individual power constraint, suboptimal25101520P,(dB)secrecy rate ((bits/symbol), 10, 9, 令, GG4, 8, 7,... | In Figures 2 and 3, we plot the second-hop secrecy rate, which is the maximum secrecy rate that our collaborative relay beamforming system can support under both total and individual relay power constraints. For the case of individual relay power constraints, we assume that the relays have the same power budgets: pi = ... | Question: What is the difference between the secrecy rates achieved under total and individual power constraints?
Answer: The highest secrecy rate is achieved under a total transmit power constraint. On the other hand, we observe that only a relatively small rate loss is experienced under individual relay power constr... | Collaborative Relay Beamforming for Secrecy | In this paper, collaborative use of relays to form a beamforming system and
provide physical-layer security is investigated. In particular,
decode-and-forward (DF) and amplify-and-forward (AF) relay beamforming designs
under total and individual relay power constraints are studied with the goal of
maximizing the secrec... |
1210.5752v1-Figure3-1.png | 1210.5752v1-Figure3-1 | Fig. 3. Performance comparison for different relay strategies atM = 4 with α = 0.1 when changingPC . | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : AF* AFO DF-XORHDF-SUP4.50.9e- DF-XORDF-SUPRA = a R. Rg= (1-a)R,40.83.50.7R=a 3R,, R= (1-a)3R,30.62.50.5R = a R, R= (1-a)R,0.41.50.310.20.50.1R= a 3R, R= (1-a)3R,%3D8101214101214(dB)(dB)(a) Average achievable rate of SU(b) Outage perfo... | In Fig. 2 and Fig. 3, we illustrate the average achievable rate of the SU and the outage performance of the primary transmission in subfigures (a) and (b), respectively, as the function of the power PC by choosing dA,C = dB,C = 0.5 and M = 4. Specifically, the rate requirements of the two PUs are symmetric, i.e., α = 0... | Question: What is the effect of the power PC on the average achievable rate of the SU and the outage performance of the primary transmission?
Answer: The average achievable rate of the SU increases as the power PC increases. However, the outage performance of the primary transmission decreases as the power PC increase... | Optimal Linear Transceiver Designs for Cognitive Two-Way Relay Networks | This paper studies a cooperative cognitive radio network where two primary
users (PUs) exchange information with the help of a secondary user (SU) that is
equipped with multiple antennas and in return, the SU superimposes its own
messages along with the primary transmission. The fundamental problem in the
considered ne... |
1412.6924v2-Figure5-1.png | 1412.6924v2-Figure5-1 | Figure 5: Total production of goods (Food + Minerals) and mean age achieved by economic agents always increased with increasing maximum contact horizons (EC7) | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 2000010020000180001600080• Total GoodsO Man Age15000140001200060100001000080004060005000400020• Total GoodsI Mean Age20005010015020025030035040050100150200250300350400Contact HorizonContact Horizon, 20000, 100, 20000, 18000, 16000, 80... | The results showed that globalization reduced prices if the increased contact horizon did not introduced additional topological barriers. Price levels for the scarce resource (minerals), were higher when the commercial horizon of the agent was smaller when barriers to trade were absent, i.e. agents could easily bridge ... | Question: What is the relationship between the total production of goods and the mean age achieved by economic agents?
Answer: The total production of goods and the mean age achieved by economic agents always increased with increasing maximum contact horizons. This is because as the contact horizon increases, the agen... | Visualizing the Invisible Hand of Markets: Simulating complex dynamic
economic interactions | In complex systems, many different parts interact in non-obvious ways.
Traditional research focuses on a few or a single aspect of the problem so as
to analyze it with the tools available. To get a better insight of phenomena
that emerge from complex interactions, we need instruments that can analyze
simultaneously com... |
1611.04842v1-Figure2-1.png | 1611.04842v1-Figure2-1 | FIG. 2: Transition probability in the recall process as a function of serial-position lag from the last word recalled. The blue curve concerns forward transitions, the yellow curve backward transitions. The asymmetry is concentrated at contiguous transitions (lag L = ±1), indicating the action of sequential recall. The... | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 0.25Forward TransitionsBackward Transitions0.2 -0.15 -0.1 -0.051015LagTransition ProbabilityLO, 0.25, Forward, Transitions, Backward, Transitions, 0.2, -, 0.15, -, 0.1, -, 0.05, 10, 15, Lag, Transition, Probability, LO FIG. 2: Transit... | It has been suggested (Kahan and Caplan, 2002) that a “discontinuity” exists between two different types of recall process. Indeed, this may be seen from Fig. 2, where only the peak for sequential transitions seems to differentiate the forward and backward curves. We are therefore required to account for the existence ... | Question: What is the relationship between the transition probability and the lag in the recall process?
Answer: The transition probability in the recall process is a function of serial-position lag from the last word recalled. The blue curve concerns forward transitions, the yellow curve backward transitions. The asy... | The Role of Word Length in Semantic Topology | A topological argument is presented concering the structure of semantic
space, based on the negative correlation between polysemy and word length. The
resulting graph structure is applied to the modeling of free-recall
experiments, resulting in predictions on the comparative values of recall
probabilities. Associative ... |
1305.3699v1-Figure1-1.png | 1305.3699v1-Figure1-1 | Fig. 1. Throughput of the OpenCL implementation for GPU and CPU devices. Execution time for modular multiplication of 3072-bit integers over 128 32-bit RNS channels. Inset: throughput in modular multiplications per millisecond according to number of OpenCL blocks; peak is reached on GPU for the nominal maximum of block... | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 10003072-bit Modular Multiplication GPU3072-bit Modular Multiplication CPU50GPU800CPU4030600201040020484096Blocks20050010001500 20002500 3000350040004500Modular Multiplications(sunwu) indyônoy(su) au, 1000, 3072-bit, Modular, Multipli... | Also [2] contains a comparative study of innovative variant implementations of RNS-based modular arithmetic in view of RSA encryption and decryption. Our experimental OpenCL implementation provides results on mid-range hardware along the lines of this previous body of work on high-range devices. As summarized in Figure... | Question: What is the relationship between the number of RNS channels and the throughput of modular multiplication?
Answer: The throughput of modular multiplication increases with the number of RNS channels. This is because the number of RNS channels determines the number of threads that can be executed in parallel. W... | Secrets from the GPU | Acceleration of cryptographic applications on massively parallel computing
platforms, such as Graphics Processing Units (GPUs), becomes a real challenge
as their decreasing cost and mass production makes practical implementations
attractive. We propose a layered trusted architecture integrating random bits
generation a... |
1101.1350v3-Figure4-1.png | 1101.1350v3-Figure4-1 | Fig. 4. The optimal tradeoff achieved by the time-out algorithm lattice stack sequential decoder for several values of b. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 10°(b = 0.6, Tout= 4.5193. 4.8347, 5.8186, 6.6717100)Re10-1(6 = 0.4, Tout = 800)= 4.6414, 5.1918, 6.2903, 7.2893Re10210-3(b = 0.1, Tout = 4 × 10ª)Re4.8127, 5.6023, 7.0910, 7.9021104Outage Probability, 4 bits/s/Hz-+- Random LAST Code 4... | In our simulation, we consider a long-term static MIMO ARQ link with M = N = L = 2, T = 3 and R1 = 8 bits per channel use. The incremental redundancy LAST code is obtained as an (m, p, k) Loeliger construction (refer to [23] for a detailed description). The frame error rate and computational complexity are plotted in F... | Question: What is the optimal tradeoff achieved by the time-out algorithm lattice stack sequential decoder for several values of b?
Answer: The optimal tradeoff achieved by the time-out algorithm lattice stack sequential decoder for several values of b is shown in Fig. 4. The figure shows that the time-out algorithm l... | Time-Out Lattice Sequential Decoding for the MIMO ARQ Channel | The optimal diversity-multiplexing-delay tradeoff for the multi-input
multi-output (MIMO) automatic repeat request (ARQ) channel can be achieved
using incremental redundancy lattice space-time codes coupled with a list
decoder for joint error detection and correction. Such a decoder is based on
the minimum mean-square ... |
1606.04646v1-Figure3-1.png | 1606.04646v1-Figure3-1 | Figure 3: The landscape of supervised cost function and unsupervised cost functions (with different levels of regularizations) along random lines that pass through the ground truth solution. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 20030350- Unsupervised expected perplexity (A = 0)X Unsupervised optimal solutionSupervised expected perplexitySupervised optimal solution30015025020010015150101005050Expected perpleку (А 0)Expected perplexity (A = 30)Optimal solution... | we can make from Figure 2(a) is that the global optimal solution to (2) (i.e., the first term in (4)) coincides with the global optimal solution of the supervised learning problem. On the other hand, there is a local optimal solution, which the algorithm could easily get stuck in, as shown in the figure. We also note t... | Question: What is the relationship between the unsupervised cost function and the supervised cost function?
Answer: The unsupervised cost function is a function of the weight matrix Wd, and the supervised cost function is a function of the weight matrix Wd and the ground truth solution. The unsupervised cost function ... | Unsupervised Learning of Predictors from Unpaired Input-Output Samples | Unsupervised learning is the most challenging problem in machine learning and
especially in deep learning. Among many scenarios, we study an unsupervised
learning problem of high economic value --- learning to predict without costly
pairing of input data and corresponding labels. Part of the difficulty in this
problem ... |
1812.00098v1-Figure2-1.png | 1812.00098v1-Figure2-1 | Figure 2: The dashed orange curve shows the forecast of the proposed global LSTM with GP local model. The black vertical line marks the division between the training and prediction regions. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : TrueTrueForecast18Forecast0.1517www160.10-15-0.05-1413-12--0.05-11-10020010015020025050150250timetime(a) electricity(b) traffic, True, True, Forecast, 18, Forecast, 0.15, 17, www, 16, 0.10-, 15-, 0.05-, 14, 13-, 12-, -0.05-, 11-, 100,... | To assess the quality of the proposed model, we limit the training, sometimes artificially by pruning the data, to only one week of time series. This results in 168 observations per time series. Figures 2a-2b show that the forecasts qualitatively on the publicly available datasets electricity and traffic from the UCI d... | Question: What is the difference between the dashed orange curve and the black vertical line in the graph?
Answer: The dashed orange curve shows the forecast of the proposed global LSTM with GP local model. The black vertical line marks the division between the training and prediction regions. | Deep Factors with Gaussian Processes for Forecasting | A large collection of time series poses significant challenges for classical
and neural forecasting approaches. Classical time series models fail to fit
data well and to scale to large problems, but succeed at providing uncertainty
estimates. The converse is true for deep neural networks. In this paper, we
propose a hy... |
1808.06579v1-Figure2-1.png | 1808.06579v1-Figure2-1 | Fig. 2: Numerical results for the mean data rate per user as a function of the number of users. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 120-X-Proposed mechanism with complete informationProposed Baysian mechanism11010090706050402080100120220240260140Number of users1601802008030Mean data rate per user, 120, -X-Proposed, mechanism, with, complete, information, Proposed,... | Fig. 2 shows the mean achievable rate by the users in the scenario with complete information in which case the BSs know the rate requirement of the users and the Bayesian framework we considered. The performance of the users in the proposed mechanism approaches their performance in the case of complete information. The... | Question: What is the effect of the number of users on the mean data rate per user?
Answer: The mean data rate per user decreases in the network due to the interference that is larger when the number of users increases in the network. Moreover, the allocated resource blocks per user decreases as the number of users in... | Contract-based Incentive Mechanism for LTE over Unlicensed Channels | In this paper, a novel economic approach, based on the framework of contract
theory, is proposed for providing incentives for LTE over unlicensed channels
(LTE-U) in cellular-networks. In this model, a mobile network operator (MNO)
designs and offers a set of contracts to the users to motivate them to accept
being serv... |
1802.03830v1-Figure3-1.png | 1802.03830v1-Figure3-1 | Figure 3: Performance of stochastic algorithms with various minibatch sizes. Here C = 10. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : -local5.25.2centrale SSR-405.15.1-SSR-80SSR-100-SSR-200SSR-500@ 4.94.9SOL-404.8SOL-804.8SOL-100SOL-2004.74.7SOL-500501001500.511.52iterationsamplex104errorerrorLO, -local, 5.2, 5.2, central, e, SSR-40, 5.1, 5.1, -SSR-80, SSR-100, -SSR... | Figure 3 shows for each method the estimated F (W) over iterations (or rounds of communication) in the left plot, and over the amount of fresh samples processed (or total computation cost) in the right plot. As a reference, the error of Local and Centralized (using n = 500 samples per machine) are also given in the plo... | Question: What is the effect of minibatch size on the performance of stochastic algorithms?
Answer: The graph shows the performance of stochastic algorithms with various minibatch sizes. It can be seen that with fresh samples, stochastic algorithms are more efficient than centralized and local algorithms. However, as ... | Distributed Stochastic Multi-Task Learning with Graph Regularization | We propose methods for distributed graph-based multi-task learning that are
based on weighted averaging of messages from other machines. Uniform averaging
or diminishing stepsize in these methods would yield consensus (single task)
learning. We show how simply skewing the averaging weights or controlling the
stepsize a... |
1912.02768v1-Figure1-1.png | 1912.02768v1-Figure1-1 | Figure 1: The TV2 solution (dash-dotted line) violates the maximum principle by attaining larger ( 1116 ) and smaller (− 1116 ) values than the original function (solid line). | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 11160.5-0.51116, 11, 16, 0.5, -0.5, 11, 16 Figure 1: The TV2 solution (dash-dotted line) violates the maximum principle by attaining larger ( 1116 ) and smaller (− 1116 ) values than the original function (solid line).<image>\nPortray... | It is easy to verify that the minimum is attained at c = 1116 and u(1) = 11 16 > 1 2 . This example is illustrated in Figure 1. It is known that for some combinations of parameters TGV reconstructions coincide with those obtained with TV2 [25, 24], hence the above example also applies to TGV. Even in cases when TGV pro... | Question: Why does the TV2 solution violate the maximum principle?
Answer: The TV2 solution violates the maximum principle because it attains larger ( 1116 ) and smaller (− 1116 ) values than the original function (solid line). This is because the TV2 solution is not constrained by the maximum principle, which states ... | Total Variation Regularisation with Spatially Variable Lipschitz
Constraints | We introduce a first order Total Variation type regulariser that decomposes a
function into a part with a given Lipschitz constant (which is also allowed to
vary spatially) and a jump part. The kernel of this regulariser contains all
functions whose Lipschitz constant does not exceed a given value, hence by
locally adj... |
2006.00156v1-Figure9-1.png | 2006.00156v1-Figure9-1 | Fig. 9. Frequency responses for different wind speeds | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : Without VIC-With traditional VICWith VIC-I-- With proposed VIC6060 1I\"60Vw=7.5m/sVw=9.6m/sVw=11.5m/s59.859.859.8313459.659.659.659.4S 59.459.459.259.259.2595920252025 59202558.858.8205030405060203040602030405060Time(s)Frequency(Hz), ... | The proposed VIC is also tested under different wind speeds to demonstrate its robustness to various operating conditions. The following values for the wind speeds are tested: 7.5 m/s, 9.6 m/s and 11.5 m/s. The frequency responses, WTG rotor speed responses and active power output increment of WTG are displayed in Fig.... | Question: What is the effect of the proposed VIC on the frequency response of the system under different wind speeds?
Answer: The frequency responses of the system with and without the proposed VIC are shown in Fig. 9. It can be seen that the frequency response of the system with the proposed VIC is much smoother than... | Nonlinear Virtual Inertia Control of WTGs for Enhancing Primary
Frequency Response and Suppressing Drive-Train Torsional Oscillations | Virtual inertia controllers (VICs) for wind turbine generators (WTGs) have
been recently developed to compensate for the reduction of inertia in power
systems. However, VICs can induce low-frequency torsional oscillations of the
drive train of WTGs. This paper addresses this issue and develops a new
nonlinear VIC based... |
1306.1392v1-Figure6-1.png | 1306.1392v1-Figure6-1 | Figure 6: Plot of the statistical estimators and of the distance from ground-truth. The estimators give a β which is smaller than, but still close to the optimal one. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 0.70.045+ Cross Validation- Error from truth0.0400.60.0350.5.0.0300.0250.40.0200.30.0150.010,0.2-0.0050.0000.000.05 0.10 0.150.20 0.25Cross Validation or Error from truth (A.U.)(((+0)I-(g)I)|(g)I)so), 0.7, 0.045, +, Cross, Validation,... | For this experimental sample we don’t know the ground truth. In order to validate the use of the two estimators we reconstruct a 2k× 2k phantom, whose reconstruction at different β’s is shown in figure 4. We apply the overlapping patches functional to provide at the same time an illustration of this new method. We use ... | Question: What is the relationship between the two estimators and the distance from ground-truth?
Answer: The two estimators are shown to be close to the optimal value, and are close to each other. The error done using the estimator can be checked on image 4b and 4c. The decrease in image quality, between the optimal ... | PyHST2: an hybrid distributed code for high speed tomographic
reconstruction with iterative reconstruction and a priori knowledge
capabilities | We present the PyHST2 code which is in service at ESRF for phase-contrast and
absorption tomography. This code has been engineered to sustain the high data
flow typical of the third generation synchrotron facilities (10 terabytes per
experiment) by adopting a distributed and pipelined architecture. The code
implements,... |
1905.06443v3-Figure3-1.png | 1905.06443v3-Figure3-1 | FIG. 3. Some possible modes for failure for selecting τ for phase space reconstruction using classical methods: (a) mutual information registering false minima as suitable delay generated from a periodic Lorenz system, (b) mutual information being mostly monotonic and not having a distinct local minimum to determine τ ... | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : Delay at first minima: 51.6-1.4-ž 1.2-1.051015202530(b)1.0-0.8-0.6-0.4-0.2-51015202530(c)1.0fFolding time: p = 1/eDelay at p: 2830.8-0.6-0.4-0.2-0.0-100200300400500CorrelationMutual Info.Mutual Info., Delay, at, first, minima:, 5, 1.6... | Even with the possibility that phase space reconstruction methods for selecting τ and n may work for choosing synonymous parameters of PE, there are a few practical issues that preclude using parameters from time series reconstruction for PE. One issue stems from many of the methods (e.g. false nearest neighbors and mu... | Question: What are the possible modes for failure for selecting τ for phase space reconstruction using classical methods?
Answer: The graph shows some possible modes for failure for selecting τ for phase space reconstruction using classical methods. These include mutual information registering false minima as suitable... | On the Automatic Parameter Selection for Permutation Entropy | Permutation Entropy (PE) has been shown to be a useful tool for time series
analysis due to its low computational cost and noise robustness. This has drawn
for its successful application in many fields. Some of these include damage
detection, disease forecasting, and financial volatility analysis. However, to
successfu... |
1711.02344v1-Figure6-1.png | 1711.02344v1-Figure6-1 | Figure 6. Transmitted signal in the single channel transmission and comparison of the output signals with and without switching for Example 2. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 20Output signal*2.5Output signal*2.5 with switching15 E........ Transmitted Signal10-10--15-201015202530TimeSignals, 20, Output, signal*2.5, Output, signal*2.5, with, switching, 15, E, ........, Transmitted, Signal, 10, -10-, -15, -20... | 𝐵 → 𝐴 are switched periodically in sequence for durations of 10 seconds. The output signal at the receiver end is shown in Fig. 6 (____ Output signal *2.5 with switching); on the same figure, the output signal of connections 𝐴𝐵 and 𝐵𝐴 which appear in Fig. 5 (____ Output signal*2.5) is replotted. As it is expected... | Question: What is the difference between the output signals with and without switching?
Answer: The output signal with switching is shown in Figure 6. It is clear that there is a difference between the direct communication with two lines without switching and communication by switching with a single transmission line.... | Use of Commutativity for Cryptology in Secret Communication | Commutativity of subsystems in cascade connected forms to form larger systems
gets worthy to improve noise disturbance properties, stability, robustness and
many other properties in system design. In this paper, another benefit of
commutativity property is indicated, as far as the author knowledge for the
first time, a... |
1903.00729v1-Figure4-1.png | 1903.00729v1-Figure4-1 | Fig. 4: Plots of fast-to-slow ratio F2S = f astBatcℎSizeslowBatcℎSize of hashing and CMS update phases for consecutive batches and for small (left) and medum (right) sketches. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 1030F2S - HashF25 - Hash25F2S - InsertF25 - InsertNumber of batchesNumber of batches(a) small on Arch-3(b) medium on Arch-3Fast-to-slow ratioO 00 O +N O96192Fast-to-slow ratio32, 10, 30, F2S, -, Hash, F25, -, Hash, 25, F2S, -, Insert,... | To utilize the heterogeneous cores on Arch-3, we applied the smart load distribution described in Section 5. We pair each slow core with a fast one, virtually divide each batch into two parts, and make the slow core always run on smaller part. As mentioned before, for each batch, we dynamically adjust the load distribu... | Question: What is the relationship between the fast-to-slow ratio and the performance of the hashing and CMS update phases?
Answer: The graph shows that the best F2S ratio for hashing is 4-to-1, while for CMS updates, a 1.8-to-1 division is the best. This is because the hashing phase is more compute-intensive, while t... | One Table to Count Them All: Parallel Frequency Estimation on
Single-Board Computers | Sketches are probabilistic data structures that can provide approximate
results within mathematically proven error bounds while using orders of
magnitude less memory than traditional approaches. They are tailored for
streaming data analysis on architectures even with limited memory such as
single-board computers that a... |
1902.06371v1-Figure5.1-1.png | 1902.06371v1-Figure5.1-1 | Fig. 5.1: Experimental results in the campus environment | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : Average Cost vs Traffic RateThroughput vs Traffic Rate252PROPIETMEED-OVNNEED-DURREAPER_6- REAPER_72-REAFER 40REOFER 24-REAPER 920--REAPERREAPER_19REAPER 21.515- -100.524Rate of Traffic(KD's)1010Rate of Traffic(KINS(a) Average cost(b) ... | Average Path Length (Cost). All four versions of REAPER have substantially lower average path lengths (Fig. 5.1(a)). As expected, PROPHET has the highest number of transmissions (about 22× that of REAPER at the heaviest load), since the other two protocols globally optimize their respective metrics. Further, REAPER exp... | Question: Which protocol has the lowest average path length?
Answer: The graph shows that the REAPER protocol has the lowest average path length. This is because the REAPER protocol uses a small world routing algorithm, which allows it to find paths that are shorter than the paths found by other protocols. | Achieving Throughput via Fine-Grained Path Planning in Small World DTNs | We explore the benefits of using fine-grained statistics in small world DTNs
to achieve high throughput without the aid of external infrastructure. We first
design an empirical node-pair inter-contacts model that predicts meetings
within a time frame of suitable length, typically of the order of days, with a
probabilit... |
2001.01388v1-Figure9-1.png | 2001.01388v1-Figure9-1 | Fig. 9: Social welfare versus α in the case with multiple entrants and fixed αβ. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 0.35e = 0.2a 0.3eß = 0.50.1640.3-0. 162ag=0.2-ag= 0.30.25a=0.50.160.20.1580.20.40.60.80.20.40.60.81(a) B = 1,W0.2(b) B = 1,W = 5Social welfareSocial welfare, 0.35, e, =, 0.2, a, 0.3, eß, =, 0.5, 0.164, 0.3-, 0., 162, ag=0.2, -ag=, 0.3... | Next we show how social welfare changes with α when αβ is fixed in the case with multiple entrant SPs. Results are shown in Fig. 9. We can see in Fig. 9(a) that when W is relatively small, social welfare first decreases then increases with α. That is because there is some welfare loss when adding a small amount of unli... | Question: What is the relationship between social welfare and α in the case with multiple entrants and fixed αβ?
Answer: The graph shows that social welfare first decreases then increases with α when W is relatively small. This is because there is some welfare loss when adding a small amount of unlicensed spectrum to ... | Price Competition with LTE-U and WiFi | LTE-U is an extension of the Long Term Evolution (LTE) standard for operation
in unlicensed spectrum. LTE-U differs from WiFi, the predominant technology
used in unlicensed spectrum in that it utilizes a duty cycle mode for accessing
the spectrum and allows for a more seamless integration with LTE deployments in
licens... |
1708.02983v1-Figure10-1.png | 1708.02983v1-Figure10-1 | Figure 10: e bene t of packed layer comes from reduced communication latency and continuous memory access. Since this is Sync SGD, the red triangles and blue squares should be at identical heights. e reason for different heights is that a di erent random number generator seed is used for the two runs. e example used Sy... | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 1088 KNL cores - Cifar dataset.0.80.60.40.2+Single Long Layer.* Multi Short Layers.100200300400 500Time (s).Accuracy., 1088, KNL, cores, -, Cifar, dataset., 0.8, 0.6, 0.4, 0.2, +Single, Long, Layer., *, Multi, Short, Layers., 100, 200... | 5.2 Single-Layer Communication Current deep learning systems [11] allocate noncontiguous memory for di erent layers of the neural networks. ey also conduct multiple rounds of communication for di erent layers. We allocate the neural networks in a contiguous way and pack all the layers together and conduct one communica... | Question: Why does the graph show that the single layer communication method is faster than the multi short layers method?
Answer: The graph shows that the single layer communication method is faster than the multi short layers method because the single layer communication method reduces the communication latency and ... | Scaling Deep Learning on GPU and Knights Landing clusters | The speed of deep neural networks training has become a big bottleneck of
deep learning research and development. For example, training GoogleNet by
ImageNet dataset on one Nvidia K20 GPU needs 21 days. To speed up the training
process, the current deep learning systems heavily rely on the hardware
accelerators. Howeve... |
1905.08894v2-Figure4-1.png | 1905.08894v2-Figure4-1 | Figure 4. Gaussian (left) and coherent (right) models: decay of the relative error in time for s = 1 (regular and Gaussian Kaczmarz) and s = 250 (block methods). | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : 10- -randomized Kaczmarz-randomized KaczmarzGaussian Kaczmarz-block KaczmarzGaussian Kaczmarz--block Kaczmarz...o.. block Gaussian Kaczmarz-... r a uAu* D..0. block Gaussian Kaczmarz10103100.511.522.533.54.5TimeTimeRelative error....R... | The former observation, that one-dimensional Gaussian sketching is beneficial for the coherent model, can be heuristically understood geometrically in the following way. In the coherent model all the rows Ai of the matrix A are approximately co-linear, so iterative projections follow roughly the same direction and do n... | Question: What is the difference between the Gaussian and coherent models?
Answer: The Gaussian model is a model of random matrices where the rows are independent and identically distributed Gaussian vectors. The coherent model is a model of random matrices where the rows are approximately co-linear.
Question: What i... | On block Gaussian sketching for the Kaczmarz method | The Kaczmarz algorithm is one of the most popular methods for solving
large-scale over-determined linear systems due to its simplicity and
computational efficiency. This method can be viewed as a special instance of a
more general class of sketch and project methods. Recently, a block Gaussian
version was proposed that... |
2001.05578v1-Figure5-1.png | 2001.05578v1-Figure5-1 | Figure 5. Retrieval error rate by the percentage of POTD database. | [
{
"from": "human",
"value": "OCR extracted text list, separated by ', ' : VLLJAAVSEC-LIJA0.90.80.70.60.50.4 -0.30.20.1-102030405060708090 100PercentageError Rate, VLLJA, AVSEC-LIJA, 0.9, 0.8, 0.7, 0.6, 0.5, 0.4, -, 0.3, 0.2, 0.1-, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, Percentage, Error, Rate Figure 5. Re... | is around or more than 1000, we consider removing irrelevant words from all modalities. Results of the error rate of POTD are shown in Fig. 5. The figure depicts the retrieval errors averaged over all testing queries. A better performance is equivalent to lower error rateCurves are closer to the bottom left corner. Fro... | Question: What is the relationship between the retrieval error rate and the percentage of POTD database?
Answer: The retrieval error rate decreases as the percentage of POTD database increases. This is because the more data we have, the more likely we are to find a relevant image for a given query. | VSEC-LDA: Boosting Topic Modeling with Embedded Vocabulary Selection | Topic modeling has found wide application in many problems where latent
structures of the data are crucial for typical inference tasks. When applying a
topic model, a relatively standard pre-processing step is to first build a
vocabulary of frequent words. Such a general pre-processing step is often
independent of the ... |
End of preview. Expand in Data Studio
README.md exists but content is empty.
- Downloads last month
- 33